United States Industrial Environmental Research EPA-600/7-80-080
Environmental Protection Laboratory April 1980
Agency Research Triangle Park NC 27711
A Case Study in the Use
of Ambient Data for
Source Assessment
Interagency
Energy/Environment
R&D Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/7-80-080
April 1980
A Case Study in the Use of
Ambient Data for Source Assessment
by
Edward T. Brookman and John E. Yocum
TRC - The Research Corporation of New England
125 Silas Deane Highway
Wethersfield, Connecticut 06109
Contract No. 68-02-2615
Task No. 18
Program Element No. INE623
EPA Project Officer: John 0. Milliken
Industrial Environmental Research Laboratory
Office of Environmental Engineering and Technology
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
DISCLAIMER
This report was furnished to the Environmental Protection Agency by TRC -
THE RESEARCH CORPORATION of New England, Wethersfield, Connecticut, in
fulfillment of Contract No. 68-02-2615, Task I 18. The contents of this report
are reproduced herein as received from the contractor. The opinions, findings,
and conclusions expressed are those of the author and not necessarily those of
the Environmental Protection Agency. Mention of company or product names is
not to be considered as an endorsement by the Environmental Protection Agency.
-m-
-------
ABSTRACT
A common objective of regional environmental management is to determine
what sources of pollution are the principal determinants of environmental
quality in a given area. This report is a case history for such an
environmental management study that was conducted for the Allegheny County
Health Department, Bureau of Air Pollution Control in Pittsburgh, Pennsyl-
vania. The results of this work were used by Allegheny County as part of their
contribution to the State Implementation Plan for achieving air quality
standards for total suspended particulate matter (TSP). The techniques that
were utilized in this ambient-correlation study include: (1) analysis of
present air quality and trends; (2) log-normal distributions; (3) relative
frequency of TSP levels; (4) monthly variations in TSP levels; (5) weekday/
weekend analysis; (6) wet day/dry day analysis; (7) analysis of pollution
roses; (8) wind frequency analysis; (9) isopleth maps; (10) contribution of
steel plant emissions by modeling; and (11) particulate identification
analysis. The integrated application of these techniques to determine the
background traditional and nontraditional components of the ambient TSP levels
is described. The results of this environmental management study include
estimates of the relative source strengths of particulates, the relative
impacts of the sources, and the level of confidence of these results.
-iv-
-------
CONTENTS
Disclaimer iii
Abstract iv
1. Introduction . .. . . . 1
2. Description of Environmental Management Study . .... .. . 3
3. Methodology Used in Management Study 5
Discussion of Individual Analytical Techniques ...... 5
Present Situation and Trends 5
Compliance History ...... 7
Air Quality Trends 7
Evaluation of Sampling Sites 8
Log-Normal Plots 8
Relative Frequency of TSP Levels 10
Monthly Variation in TSP Levels . . .10
Weekday/Weekend Analysis 10
Wet Day/Dry Day Analysis 13
Pollution Roses 14
Wind Frequency Analysis 14
Isopleth Maps of Particulate Patterns . . 16
Contribution of Steel Plant
Emissions to Selected Monitors ...... 16
Particulate Identification Analysis ........... 1.6
Integrated Application of Techniques . ... 18
Background Component 19
Traditional and Non-Traditional Components 20
4. Results of the Management Study 22
5. Critique of the Methodology Used in the Management Study . . 24
-------
FIGURES
Number Page
1. Hi-Vol Monitoring Stations for Allegheny County ........ 6
2. Example of Log-Normal Plot for TSP Data 9
3. Example of Relative Frequency Plot for TSP Data 11
4. Example of Seasonal Variation Plot for Two Sampling Sites . 12
5. Example of Integrated Pollution Rose for TSP Data 15
i 6. Example of Isopleth Map Showing Pattern of TSP Levels
in ug/m3 17
VI
-------
TABLES
Number Page
1. Annual Arithmetic Average Component Breakdown For
Particulate Matter in Allegheny County (1975-1977) .... 23
vii
-------
SECTION 1
INTRODUCTION
A frequent task In environmental management for a given region or locality
is to determine what sources of pollution are the principal determinants of
environmental quality in that area. If this determination can be established
with a reasonable degree of certainty, then the regulatory decisions necessary
to attain compliance with environmental quality goals for the area can be made
with minimal risk. In most cases, however, the determination of relative
contributions from the multitude of sources is complicated, especially where
the pollutant is produced from natural or secondary sources.
Therefore, one objective of the environmental manager is to understand
the relationship between ambient environmental quality and the sources which
determine that quality. The Environmental Protection Agency (EPA) is
interested in analyzing the various methodologies that can be employed in
accomplishing this objective. Because any one methodology may not be
universally suited to the variety of environmental management problems
encountered, it is desired to examine a typical study of ambient-source
correlation where the integrated use of several techniques is employed. This
report describes an environmental management study that TRC-THE RESEARCH
CORPORATION of New England conducted for the Allegheny County Health
Department Bureau of Air Pollution Control (BAPC) in Pittsburgh, Pennsylvania,
in which a number of analytical techniques were used to estimate the relative
contributions from various sources of total suspended particulate (TSP) within
the County.
This report discusses this management study in the following manner:
o Section 2 describes the scope of the environmental management study,
the results expected, and the users of the study outputs.
o Section 3 describes the methodology used in the management study,
including:
- present situation and trends
- log-normal TSP distribution analysis
- relative frequency of TSP levels
- monthly variations in TSP levels
- weekday/weekend analysis
- wet day/dry day analysis
- pollution roses
-1-
-------
- wind frequency analysis
- isopleth maps
- contribution of steel plant emissions to selected monitors by
modeling
- particulate identification analysis
The integrated application of these techniques to determine the
background, traditional, and non-traditional components of the
ambient TSP levels is then discussed.
Section 4 describes the results of the management study, including
the sources of particulates, the relative impacts of the sources,
and the level of confidence of the results.
Section 5 presents a critique of the methodology used in the
management study and gives recommendations that would benefit future
studies.
-2-
-------
SECTION 2
DESCRIPTION OF ENVIRONMENTAL MANAGEMENT STUDY
Allegheny County, Pennsylvania, has been identified by EPA as a
nonattainment area for TSP that, under the Clean Air Act of 1977, must submit a
revised State Implementation Plan (SIP) to EPA describing how the National
Ambient Air Quality Standards (NAAQS'S) for particulate matter will be
achieved by 1982. TRC was contracted by the BAPC to organize and manage an
effort to develop strategies for the control of traditional and non-
traditional sources of suspended particulate matter in Allegheny County.
Several organizations participated in various parts of the project, including
Carnegie Mellon University, U.S. Steel's Research Center, Energy Impact
Associates, and Materials Consultants & Laboratories.
As part of this management study, several years of ambient TSP data as
measured by Hi-Volume Samplers in the County were to be analyzed. The
objective of the analysis was to use the available data to determine the
relative contributions of industrial sources (traditional and non-
traditional), non-traditional sources (such as road dust) and background at
selected sites in the County that were considered to be indicative of the
particulate problem and that could help define strategies for particulate
control.
Sulfur dioxide and other specific gaseous pollutants can usually be
traced to specific sources because they retain their identity from the emission
source, to the point of measurement in the ambient atmosphere. Analogous
tracking of particulate matter from source to receptor is much more difficult,
and often not achievable in practice. Particulate matter is ubiquitous. The
particles that are captured on the filter of a Hi-Vol can come from a number of
sources and arrive at the point of capture after following innumerable routes.
Furthermore, the range of physical and chemical properties of particulates are
almost limitless, and the use of these properties is not a straightforward
means of identifying the sources and routes followed by the collected
particles. Therefore, a methodology was developed to determine the relative
contributions to total particulate matter measured by broad source or route
classifications of particles. Four classifications were considered as
follows:
-3-
-------
Classification
Traditional Sources (Industrial)
In-Plant Non-Traditional Sources
(Industrial)
3. Non-Traditional Sources
(Non-Industrial)
4. Background Material
Definitions & Examples
"Virgin" (non-resuspended) mater-
ial arriving at the sampling point
directly from point and process
fugitive sources within a plant
complex.
Fugitive dust from wind blown
storage piles and materials handl-
ing and resuspended dust from
traffic on dusty plant roads.
Dust from construction and demo-
lition activities. Re-entrained
dust from road traffic, play-
grounds, parking lots, etc.
Particulate matter of both natural
and anthropogenic origin advected
from points outside the County and
over which the County has no
control.
-4-
-------
SECTION 3
METHODOLOGY USED IN MANAGEMENT STUDY
The primary body of data available for the study was that amassed by the
BAPC for the years 1975 through 1977. During these years, particulate data
were collected essentially continuously at twenty-four sampling stations in
the County (Figure 1). These particulate data, together with meteorological
data from the Greater Pittsburgh Airport, were used in statistical analyses and
manipulations which provided the principal basis for determining the relative
contributions of background, non-traditional, and traditional sources of
particulates.
DISCUSSION OF INDIVIDUAL ANALYTICAL TECHNIQUES
The individual analytical techniques used in the management study were:
o Present situation and trends
o Log-normal TSP distribution plots
o Relative frequency of TSP levels
o Monthly variations in.TSP levels
o Weekday/weekend analysis
o Wet day/dry day analysis
o Pollution roses
o Wind frequency analysis
o Isopleth maps of particulate patterns
o Contribution .of steel plant emissions to selected monitors using
modeling techniques
o Particulate identification analysis
The above methods are described in the following subsections.
Present Situation and Trends
The development of logical and effective strategies for further control
of particulate matter to meet NAAQS's in Allegheny County requires a thorough
understanding of trends in both emissions and ambient concentrations of
particulate matter. Furthermore, the siting of particulate monitoring
stations has an important bearing on the levels of particulate matter measured.
Therefore, this analysis consisted of three sub-analyses: a compliance
history; an analysis of present ambient particulate levels; and an evaluation
of the present particulate air monitoring stations.
-5-
-------
Figure 1. Hi-Vol Montioring Stations for Allegheny County.
-6-
-------
Compliance History—
As a first step in developing strategies for the further control of
participate matter, it is mandatory to determine how effective present
regulations have been and are likely to be in reducing particulate emissions.
Allegheny County has had in force since 1972 a set of regulations for the
control of particulate emissions that are as stringent as any in the country.1
Furthermore, those emission regulations that apply to steel making operations,
the principal heavy industry in the County, represent levels of emission
control that equal or exceed New Source Performance Standards for applicable
processes.
A compliance history was prepared to show the relative amounts of
particulate control that have been achieved roughly to the present (1976 is the
base year for the emission inventory) and that are expected to be achieved by
1982 by continued enforcement of the requirements of Article XVIII and various
consent decrees. The beginning year for this analysis was set as 1971, just
before the enactment of Article XVIII.
The changes in emissions over the period 1971-1982 have been and will be
the result of installing control devices, varying fuel mixes, varying
production levels, phasing out of old processes, the use of new processes,
improved maintenance practices, and other changes. The work performed to
assemble the compliance history was performed by BAPC personnel.
The compliance history showed that there was a 65 percent reduction in
particulate emissions between 1971 and 1976 and that the anticipated
reductions in particulate emissions between 1976 and 1982 will be an additional
49 percent.
Air Quality Trends—
Hi-Vol sampling is currently being carried out at 24 locations in
Allegheny County. Continuous sampling at each of these stations every 3 or 6
days has been in progress since 1975. Eight of these stations have been in
continuous use since 1970. The yearly arithmetic and geometric means at each
site were computed and plotted in order to determine whether any large
anomalies exist in the air quality trends at any of the monitoring stations.
This analysis technique can provide the following types of information:
o A large change in the yearly TSP levels at only one site can indicate
a local source starting up or shutting down. Such a local source
could be a construction site, strip mine, or small industrial
source.
lAllegheny County Health Department Rules and Regulations Article XVIII, June
1972 and Amendments 1978.
-7-
-------
o A large change in the yearly TSP levels at several sites in a large
geographical region can indicate a major TSP source starting up or
shutting down or undergoing a major change in operations. Such a
major source could be a steel plant.
o Gradual changes in the yearly TSP levels at a particular site or
several sites can indicate the effectiveness of implemented control
measures.
Evaluation of Sampling Sites—
To properly assess the representativeness of the particulate sampling
stations to human exposure, an analysis of each site should be performed.
These anlayses should examine monitor placement, location of local sources,
type of neighborhood, etc. Such an analysis was performed for fifteen of the
Allegheny County sampling sites. In addition, micro-inventory site summaries
were performed for an additional eight sampling sites.
Another way of examining the representativeness of monitoring sites is by
using the Standard Air Monitoring Work Group (SAMWG) guidelines for monitor
placement.2 These guidelines take into account the monitor height, distance
from roads, freedom from airflow obstructions, etc. This was done for all of
the Allegheny County monitoring stations.
Based on the SAMWG guidelines, observations were made as to the
representativeness of the monitoring station sites chosen for use in Allegheny
County. Several of the sites were felt to be non-representative of the general
TSP levels within the County and were, therefore, not used in the subsequent
attainment analysis.
Log-Normal Plots
An analysis method that can be useful in determining the presence of local
sources is the log-normal distribution plot in which the percentage of readings
over a certian TSP level is plotted versus that TSP level. If a site is subject
to large-scale or general influences (i.e., many sources), then its TSP
observations should be log-normally distributed. If there is a major local
influence, such as a stack or a strong area source in a specific direction from
the sampling site or some other source of extreme impact, then the data will
either not exhibit log-normality or will deviate from it at the plot extremes.
The facilities of the U.S. Steel Research Center were used to produce
log-normal plots for each of the 24 monitoring sites for each of the three
study years. An example of such a plot is given in Figure 2.
2Air Quality Surveillance Network Design and Siting for State Implementation
Plan (SIP) Monitoring.
-8-
-------
1000
I I \ t t t t
10 .50 2.0 10.0 20.0 60.0 90.0
99.0 99.90 99.99
Figure 2. Example of Log-Normal Plot for TSP Data.
(1976 Data For Station 8702-Liberty Boro)
-9-
-------
These plots showed that data from all stations appear to be reasonably
log-normally distributed which implies that all stations are subject to large-
scale influences. This result was not expected since several stations are near
to strong point or area sources. The log-normal distribution of data for one
station which recently has been influenced' by a nearby trucking operation
showed only minor deviations from log-normality.
Relative Frequency of TSP Levels
As in the case of log-normal plots, a clue as to whether a monitoring site
is affected by a single source can be obtained by inspection of the
concentration frequency distribution for that monitor. If the distribution
shows extreme values of much larger concentration than the mass of data, then a
directionally dependent effect on that monitor or other similar concentrated
causes for extreme values may be suspected. Again, with the help of U.S.
Steel's Research Center, graphs of relative frequency were prepared for each
site for each year of the study period. An example of this type of graph is
shown in Figure 3.
In reviewing the frequency distribution graphs, a qualitative judgement
was made as to any evidence that extreme values might influence the shape of
the distribution curve. We were not able to explain why, over the period of
1975-1977, extreme values appear to have an increasing influence on the pattern
of frequency distributions.
Monthly Variations in TSP Levels
The three years of TSP data were averaged by month and plotted to show the
monthly and seasonal variations that exist at each of the sites. An abnormally
high winter level can be an indication of heavy traffic influence due to the
combination of longer morning inversion periods, cars idling while cold, and
road sanding/salting operations. A high summer level can be an indication of
activity in agriculture or the increased use of school playgrounds.
A variation of this analysis was also performed where the monthly average
plots were grouped for stations that are similar in character or in close
proximity. This helps to determine whether these groups of stations are
influenced by the same source regime. An example of this type of plot is given
in Figure 4. Note how the two stations exhibit similar seasonal TSP patterns,
indicating the same general source regime; while one station has a higher
overall TSP level, indicating a local source influence. These two stations are
within one kilometer of each other.
Weekday/Weekend Analysis
The arithmetic means for weekday and weekend periods as well as Saturday
and Sunday individually were computed for each site for each study year and for
-10-
-------
113 OBSERVATIONS
LOWER
CELL
LIMIT
450.00
420.00
390.00
360.00
330.00
300.00
270.00
240.00
210.00
180.00
150.00
120.00
90.00
60.00
30..00
0.00
MINIMUM • 3.6000000E+01
MAXIMUM « 4.3600000E+02
RELATIVE FREQUENCY
CELL
MID-
POINT
435.00;
405.00
375.00
345.00
315.00
285.00
255.00
225.00
195.00
165.00
135.00
105.00
75.00
45.00
13.00
MEAN
STD. OEV.
GEO. MEAN
STD. OEV. OF LN.
1.1128319E+02
5.9511774E-K11
99.57057
.4615998
CUMULATIVE FREQUENCY
FREQ.
0
I
0
Q
1
0
0
I
3
5
9
21
24
32
16
0
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.3 0.9 1.0
xsexxx
xx xxx xa xxx
XM XXXXXXXX XX
XX XXXXXXXX XXXXX
XXXXXXX »
Figure 3. Example of Relative Frequency Plot for TSP Data.
(1977 Data for Station 5801-County Office Building)
-11-
-------
160
150 -
•• 140 -
0»
E 120
I
« 110
100 -
90
i i i
7570 DUQUESNE 2
7502 OUQUESNE 1
t i t i t t
JFMAMJAMSONOJ
MONTH
Figure 4. Example of Seasonal Variation Plot for Two
Sampling Sites. (1975-1977 Data)
-12-
-------
all three study years combined. With only one exception, all of the stations
exhibited lower particulate levels on weekends than on weekdays. One might
expect that this difference would be caused by a combination of reduced
industrial activity (traditional sources) and reduced traffic (non-traditional
sources) on weekends. In the Monongahela Valley, which is dominated by steel
making operations that operate at a relatively constant level every day, one
would expect that the weekend contributions from traditional sources would
equal those during weekdays.
Traffic data could not be obtained at any of the County monitoring
stations, but some information on traffic patterns was obtained for the state
in general. These results were surprising in that they showed significantly
more automobile traffic on weekends than on weekdays. If the vehicular use in
Allegheny County is similar to that state-wide, the average daily traffic
volume appears to have a negative effect on particulate concentrations.
However, assuming that traffic resuspends particulate matter that can affect
Hi-Vol readings, there are at least two factors that could make the traffic
data not consistent with the particulate data:
1. Weekend contributions from plant generated fugitive dust may be
significantly less than for weekdays.
2. Weekend driving patterns are different than those for weekdays.
During weekends, driving is done in the middle of the day when dispersion
conditions are at their best. During weekdays, peak driving is done in the
early morning (0630-0900) and late afternoon (1500-1800) when dispersion
conditions may be poor, thus keeping traffic-suspended particulate matter in
the vicinity of the point of generation in a relatively undiluted condition.
We were not able to obtain specific emission inventory data on a
weekday/weekend basis, but qualitative reports from the principal steel
companies in the County showed that, while steel production was relatively
constant throughout the week, certain operations such as shipping, loading,
and unloading were at reduced levels over weekends. Such operations are
important non-traditional particulate sources.
Wet Day/Dry Day Analysis
In performing this analysis, it was assumed that on days with sufficient
snow coyer, with greater than 0.5 centimeters of precipitation, or following
days with greater than 0.5 centimeters of precipitation, the principal
contributions to the TSP levels would be traditional sources, home heating,
vehicle exhaust, and material transported from outside the County. The rain
and snow would suppress the majority of the local fugitive and reentrained
dust. These days were defined as "wet" days and all other days as "dry" days.
The meteorological data for the three-year study period were examined and the
TSP levels at each site on wet days were then averaged and compared to the
average TSP levels on the dry days. The results showed that on wet days total
-13-
-------
TSP was from 14 to almost 50 percent less than on dry days. These differences
give a first approximation of the non-traditional source component at each of
the sampling stations.
Pollution Roses
Pollution roses, which depict the average TSP concentration for various
wind directions, were constructed using the computer facilities of the U.S.
Steel Research Center for each site for each of the study years. They were
based on the sixteen compass points for a wind persistence factor >0.71.3 For
each site, two plots were constructed for each year: one for wind" speeds of
0-3.5 m/sec and one for wind speeds of 3.5-20 m/sec.
To facilitate subsequent analyses, the data were then combined for all
wind speeds and a persistence factor >0.71 and plotted in accordance with the
eight cardinal compass directions. An example of this type of plot is
presented in Figure 5.
These roses are useful in determining if high TSP levels are associated
with a particular wind direction or directions. They are capable of showing
the following types of influence from sources:
o Lack of any specific directional effect of sources on background
stations.
o Diffuse influence of distant industrial complexes.
o Combination of diffuse influence from distant sources and nearby
sources.
o Influence of nearby large sources.
Wind Frequency Analysis
As an aid in performing the analysis, the percent of time the wind blows
from a particular direction (persistence >0.71) was determined from the
Greater Pittsburgh Airport wind data. This wind frequency can be combined with
the pollution rose information to determine the TSP level contribution from a
particular compass sector.
3Wind persistence factor is defined as the ratio of vector average wind to the
average wind speed over the 24 hour Hi-Vol sampling period. A factor near 1.0
indicates a wind that blows consistently from one direction during the entire
sampling period. A persistence factor >0.71 is equivalent to an hourly wind
direction deviation of 45°.
-14-
-------
200 ug/m3
Figure 5. Example of Integrated Pollution Rose for TSP Data.
(1975-1977 Data for Station 8601 - Glair-ton)
-15-
-------
Isopleth Maps of Participate Patterns
Data for generating participate isopleths, or lines of constant concen-
tration, were developed with the help of the U.S. Steel Research Center. The
TSP data were stratified into periods when wind conditions over the sampling
wind period met certain criteria: wind direction within one of the eight
compass sectors on days with persistence >0.71 and within two wind speed
classes (0-3.5 m/sec and 3.5-20 m/sec). The average TSP concentrations for
each site, for each wind direction, and for the two wind speed regimes were
then combined, based on the number of observations, to give the TSP
concentrations for each wind direction for all wind speeds. These values were
then used to form isopleth maps representing particulate concentrations
associated with winds from each of the compass directions. An example of such
a map is presented in Figure 6. Recognizing the general locations of the
industrial and urban sources, these isopleth patterns show the effect of wind
"smearing" of these emissions downwind of the sources.
Contribution of Steel Plant Emissions to Selected Monitors
Under the sponsorship of the steel companies with plants in Allegheny
County, Energy Impact Associates (EIA) of Pittsburgh, Pennsylvania, carried
out an indepth assessment of the particulate impact of each of the steel making
facilities on selected monitors. The work consisted of the following elements:
o Development of a detailed in-plant particulate source inventory
including emission estimates for fugitive sources.
o Development and application of a model for predicting ambient
particulate concentrations. The model accounts for particulate loss.
through deposition.
EIA's calculations were in terms of the yearly arithmetic averages for 1976 and
1982, utilizing Greater Pittsburgh Airport winds. EIA stratified the impact in
terms of point, process fugitive, and non-traditional sources from within the
plant area. We then adjusted the 1976 averages to the three-year average data
base on which our analysis was based by simple ratioing.
Particulate Identification Analysis
A field sampling program was conducted which collected TSP samples using
membrane type Hi-Vol filters at 12 different Hi-Vol sites in the County. A
total of 50 samples was collected including 33 ambient samples (21 day, 12
night), 15 control samples, and 2 special test filters. The sampling was
conducted under carefully specified wind and stability conditions for each
sampling group by Denardo and McFarland Weather Service, Inc.
Each ambient and reference filter was analyzed by computer-controlled
Scanning Electron Microscopy/Energy Dispersion X-Ray Analysis(SEM/EDAX). This
-16-
-------
Figure 6. Example of Isopleth Map Showing Pattern of TSP Levels
in ug/m3. Example is for Southeast Winds.
-17-
-------
technique classifies particles into various types. The particles are then
classified as being from traditional or non-traditional sources by comparing
chemistries, sizes, and shapes of particles on the ambient filters with
particles from control samples taken from various industrial processes, as
well as from known sources of non-traditional particles such as clay, sand, and
street dust. This analysis was performed by Materials Consultants &
Laboratories, Inc. (MCL), Monroeville, Pennsylvania. They provided TRC with
the following information:
o The distribution of particles by size in terms of 48 different
particle chemistries (plus a miscellaneous category) for ambient
filters collected under specified meteorological conditions and for
17 reference samples representing various traditional and non-
traditional sources.
o The percentage by weight of large particles (>15 urn) and the
chemistries of such particles.
o A description of the 49 particle chemistries in terms of most
probable sources, both traditional and non-traditional.
TRC carried the MCL description of particle chemistries one step further
by estimating the relative percentages of traditional, in-plant non-trad-
itional, and urban non-traditional particulate making up each of the particle
chemistries. These estimates were then applied to the results of the filter
analyses made by MCL. The estimated percentages were multiplied by the weight
percentages for each of the particle chemistries making up essentially all of
the total weight of the filter catch. The contributions were then summed and
the resulting percentage breakdown of traditional, in-plant non-traditional,
and urban non-traditional components were tabulated for each sample.
INTEGRATED APPLICATION OF TECHNIQUES
Using the results of the previously described analyses along with other
published information it was possible to determine the approximate background
TSP level for the County as well as the approximate non-traditional and
traditional components of the mean TSP level at each monitoring site. The
following subsections describe the procedures used.
Before discussing the relative contribution of various components of the
TSP measurement, consideration must be given as to which of the two standards
is the more restrictive: annual geometric mean of 75 ug/m3 or the short term
standard of no more than one 24-hour value exceeding 260 ug/m3. The method
used consisted of computing the standard geometric deviation (SGD) for the data
from each station and comparing this with the SGD for the standards. The SGD
of the primary standards is the slope of the line of a log-probability plot
passing through both the 24-hour and annual standard. If the SGD of the
monitor data is greater than the SGD of the standard, the short-term standard
is more restrictive. Conversely, in the SGD of the monitor is less than that
for the standards, the annual standard is more restrictive.
-18-
-------
The results of this comparison showed that, with only a few exceptions,
the annual standard was the more restrictive and was therefore used as a basis
for the further analyses.
Background Component
For purposes of developing a control strategy for particulate matter,
background concentrations must be taken into consideration. As used in the
study, "background concentration" is that portion of the measured ambient
levels of particulate matter which is not attributable to emissions from
manmade sources within the County. This is essentially material that is
transported into the study area from external sources over which the County has
no control and material generated within the County from natural and
agricultural sources. The mean background TSP level was established through
the use of an "eclectic" particulate rose and the frequency of wind direction.
To remove the impact of Allegheny County sources on TSP levels within the
County, only those wind directions not associated with County sources were
chosen at two selected monitoring sites located at the far southwest and
northeast edges of the County. The TSP concentrations at these sites were then
obtained for all days with a wind persistence factor >0.71. These were
classified by wind direction sector and the mean of each data set was
calculated. A composite (or "eclectic") particulate rose was constructed
utilizing the wind direction dependent values from the two sites associated
only with wind directions not associated with County sources. To determine a
weighted mean background level, the percent of time the wind blew from each
directional sector was used. The mean TSP level for each directional category
was multiplied by the frequency of wind from that category and the results were
summed for all eight categories in order to give the weighted mean.
The weighted values obtained were assumed to be composed of transported
material, home heating, vehicle'exhaust, reentrained road dust, tire rubber,
and natural and agricultural sources. Since the sites used are located in
relatively rural or suburban locations and due to the wind directions chosen,
the contributions from home heating and vehicles were assumed to be quite
small. Representative values for the emissions from these two sources were
chosen using published information. A background TSP level for Allegheny
County was thus established. In subsequent analyses, this background level was
applied uniformly to data at all sampling sites within the County.
The background level developed for Allegheny County is 51 ug/m3 annual
average and 48 ug/m3 annual geometric mean. These levels are high when
compared with other urban areas, but considering the strength of industrial
sources to the west and southwest of the County, this level of background is
considered reasonable. Other investigators using less rigorous techniques
have confirmed that background levels of this order should be expected for
Allegheny County.
One other background value was computed in order to provide some
information on the transported portion of the background level. It was assumed
-19-
-------
that on days with sufficient snow cover, with greater than 0.5 centimeters of
precipitation, or following days with greater than 0.5 centimeters of
precipitation, the only contributors to the Hi-Vol filter would be home
heating, vehicle exhaust, and transported material from outside the County for
"eclectic" wind directions. The rain and snow would suppress the fugitive dust
sources. The meteorological data for the three-year study period were examined
and those days meeting the above criteria and with wind persistence >0.71 were
noted. The TSP levels at the background sites were then obtained for those
particular days, sorted by wind direction, averaged and weighted as before.
The value obtained was assumed to consist of home heating, vehicle
exhaust, and material transported from outside the County. As before, using
published information, the combined effects for home heating and vehicle
exhaust were estimated. A value was thus established for the "wet" background
level, or amount due to transport, in Allegheny County. This value is 36 ug/m3
annual average.
Traditional and Non-Traditional Components
Once the background level has been determined and subtracted from the
three-year mean, the remainder will be composed of traditional and non-tradi-
tional components. For the purposes of the study, the traditional component
was assumed to be composed of industry-related sources (stack, process
fugitive, and inplant fugitives) and other fuel combustion sources (home
heating) that impact the sampler directly. The non-traditional component was
assumed to be composed of vehicle-related emissions (reentrained dust, tire
rubber, exhaust, brake linings, etc.), construction-related sources, and other
miscellaneous non-industrial sources.
To obtain the approximate component breakdown at each site, a variety of
methods was used which encompassed any or all of the analyses described
previously along with other published data calculation methods. The details of
the evaluation at each site will not be described here since they are quite
involved and would require lengthy explanations. However, the general
methods used will be discussed and some examples will be given.
Pollution roses were used extensively in the analyses of the sites. The
first step was to subtract the values of the background rose from the
site-specific rose. The remainder could then be examined in relation to the
directions of traditional sources and to the directions of local non-tradi-
tional sources (from the site evaluations) and a good indication of the
non-traditional and traditional contributions could be obtained.
The results of the EIA study were also used extensively. This study was a
rigorous treatment of the impact of steel mill sources on selected monitors and
was considered to be a fair representation of the actual conditions. A few of
EIA's results were modified where it appeared that part of their contribution
was more properly labeled "non-traditional" than "traditional." The MCL
results were used as a check of the findings of the EIA near-field modeling
study.
-20-
-------
Direct, day-by-day comparison of TSP levels at various sites resulted in
valuable information on component breakdown. This was particulary useful at
sites that were in close proximity or that were in very similar geographical
locations. For example, it was seen that for wind directions not associated
with the traditional sources, the daily TSP levels at site A were consistently
10-20 ug/m3 higher than the levels at a neighboring site B. This indicated
that the non-traditional influence was about 15 ug/m3 higher at site A.
At several sites, engineering judgements were made as to the effect of
traditional sources. For instance, it was assumed that two of the rural sites,
located in the western part of the County, would be affected identically by the
traditional sources within the County, due to their approximately equal
distance from the source locations.
The site evaluations were also of great importance since they indicated
the location and size of local fugitive dust sources such as playgrounds and
parking lots.
The other analyses, such as the seasonal variations, wet/dry comparisons,
weekday/weekend comparisons, etc, were valuable in providing indications of
whether the non-traditional or traditional sources predominated.
-21-
-------
SECTION 4
RESULTS OF THE MANAGEMENT STUDY
By the procedure outlined in the previous section, the component
breakdown at each site was determined. The breakdown was calculated both in
terms of arithmetic and geometric means. The final results, in terms of the
arithmetic means, are reproduced in Table 1.
• An examination of this table reveals the relative contributions of the
traditional and non-traditional components of the mean TSP levels at each of
the sites within Allegheny County. It can be seen that several sites, such as
the Court House (5802) and Pittsburgh Airport (4401), are predominantly
affected by traffic and other non-traditional sources, while others, such as
the Braddock stations (7102 and 7104), are dominated by traditional sources.
These sites are impacted primarily by the Edgar Thompson Works of U.S. Steel.
Several stations such as Swissvale (7004) and Hazelwood 2 (6904), are affected
fairly evenly by both traditional and non-traditional sources.
It should be noted that the values presented in Table 1 are based on a
wide variety of techniques plus a considerable amount of engineering
judgement. It is not possible to assign confidence levels in statistical terms
to the values. However, we believe that the values are an accurate assessment
of the relative contribution of various sources to particulate levels in
Allegheny County. Individual values should not be taken as firm figures. The
pollution roses were generated using airport winds and these have been shown by
EIA to vary considerable from local winds. There are numerous topographical
features in Allegheny County that have large effects on airflow patterns.
Nevertheless, the patterns of contributions as a whole are a reasonable
approximation of the true picture and TRC believes they are an adequate basis
for the development of particulate control strategies to meet air quality
standards by 1982. Furthermore, we believe that the approach used here, which
relies upon a variety of methods and a considerable amount of engineering
judgement, is a considerable improvement over the traditional approach of
using source inventories together with unvalidated models. To gain further
insight as to the exact components and their sources at each site, more
detailed studies and measurements are warranted.
-22-
-------
TABLE 1. ANNUAL ARITHMETIC AVERAGE COMPONENT BREAKDOWN FOR PARTICIPATE MATTER
IN ALLEGHENY COUNTY (1975-1977) - ug/m3
Site
Number
3001
3101
4401
4601
5602
5702
5801
5802
6201
6903
6904
6905
7004
7102
7104
7201
7502
7570
7601
3001
3601
3602
3702
3704
Site Location
Logan's Ferry
Sprfngdala
Greater Pitt. Airport
Bellevua
Murray Towers
Central Lab
County Office Bldg.
Court House
North Fayetta
Hazalwood 1
Hazalwood 2(a)
Kaufmaim
Sw1ssva1e'a'
8raddocfc(a)
North 3raddodt(a)
wall
Ouquesne V*'
Ouquesne 2
Allegheny Co. Airport
South Fayette
Cla1rton(a)
Qlassport(a)
L1berty(a)
Coursln Hallow^
3-Year
Arithmetic
Average
73
71
31
110
102
' 118
101
184
66
107
142
31
150
164
131
81
117
143
33
61
110
134.
123
110.
Badc-
Ground
51
51
51
51
51
51
51
51
51
51
51
51
51
51
51
51
51
51
51
51
51
51
51
51
Traditional
Other Non-Traditional
Stack Trad. Industrial Urban
5
5
1
5
2
2
2
2
1
3
5
2
8
7(b)
7<0
2
3(c)
8
2
1
^(d)
19(d)
15
-------
SECTION 5
CRITIQUE OF THE METHODOLOGY USED IN THE MANAGEMENT STUDY
As a means of determining the relative contributions of sources for the
development of control strategies, the methodology described in this report
has been effective. The methods used provided a thorough understanding of the
nature of the particulate problem in Allegheny County. The interpretation of
the information developed led to the isolation of the contributing sources. In
particular, the following analyses should be included in a study of this type:
o Trend analyses
o Site analyses - these must be done in order to identify local
influences that could affect the sampler data
o Monthly/seasonal analyses
o Pollution roses
o Filter analyses
The other analyses that were performed were more useful as indicators of
whether traditional or non-traditional sources predominated at any particular
site, rather than as methods to identify an individual source. These analyses
are:
o Log-normal plots
o Weekday/weekend analyses
o Relative frequency plots
o Wet day/dry day analyses
o Isopleth contours
While the methodology was effective, there are several things that should
be done to improve its usefulness in future studies. The following items
should be considered by environmental managers:
o The overall placement of particulate samplers should be reviewed in
relation to sources and human exposure, as a means of truly
representing attainment or non-attainment of air quality standards
and providing adequate geographical coverage of the entire study
area. Usually, large portions of an area are not covered by sampling
sites, while there is considerable redundancy of sampling in
so-called "hot spots."
-24-
-------
o Rather than relying on meteorological data from one location, such
as the nearest airport, detailed meteorological measurements should
be recorded simultaneously with the sampler data at each site or at
least in several different representative locations. The local
winds can vary significantly from location to location and the
determination of the local wind patterns can greatly help in source
identification.
o Continuous ISP data should be obtained at several sampler locations
(i.e., daily, rather than every third or sixth day) to allow for the
influence of traffic, as well as providing more detailed information
on TSP patterns in relation to meteorological influences and
variations in source strengths.
o Traffic flow data should be obtained concurrently with the TSP
sampling data to allow for the correlation between these two
parameters. The data should be stratified by day of the week. This
would enable the weekday/weekend analysis to be a more useful
analytical technique, and would also allow a slant distance analysis
to be performed.*
o Information on the mixing depth, or the depth of the low-lying
unstable layer of air below a stable inversion layer which limits
atmospheric dilution of pollutants, would be very useful in
establishing the relationship between inversions and TSP levels.
o More statistical analyses, such as correlation between particulate
levels and variables such as wind and other meteorological con-
ditions, could be performed on the data to establish confidence
levels.
Two other points to make here that might be of help to the environmental
manager are the use of predictive modeling and the use of particle
identification studies.
o TRC feels that the state-of-the-art of predictive modeling is
inadequate at this time to determine with precision the relative
contributions of traditional and non-traditional sources in large
urban areas. Attempting to use models in any general way based on
the currently available information could lead to erroneous con-
clusions and ill-conceived strategies for control of particulate
matter. This statement is not applicable to situations where
^Several published studies (e.g., National Assessment of the Urban Particulate
Problem. Volume I. National Assessment. Lynn, D.A., et.al., EPA-
450/3-76-024, July, 1976) indicate that there is a direct relationship between
the daily TSP concentration and average daily level of traffic and an inverse
relationship between TSP and the distance of the Hi-Vol sampler from the
traffic, measured by the slant distance.
-25-
-------
sources can be specified in detail as to configuration, mechanism of
generation, and particle size of emissions; where the model can
accept and deal with all of the above inputs plus complex terrain;
and where meteorological measurements for application of the model
have been made in the near vicinity of sources and receptors. Such a
rigorous approach was used in the special near-field study by EIA
described earlier in this report.
o In this program, particle identification techniques were used to
establish the source origins of various types of particles. While
such methods provide information on the mechanisms by which
particles are formed, they do not necessarily show the path by which
the particles arrived at the samplers. In developing air quality
control strategies, the path followed by a particle is an important
factor. For example, did the particle arrive directly from a
process, or did it deposit on the ground and become entrained?
Samples collected for particle identification must be carefully
planned, and samples under enough meteorological and source strength
conditions must be collected to be able to establish overall
contributions of specific sources to a given sampling location.
The objective of our analysis was to use available data to determine the
relative contributions of industrial sources (traditional and non-tradition-
al), non-traditional sources, and background at selected sites in the County
that we believed to be indicative of the particulate problem and could help
lead the way to strategies for particulate control. The methodology described
in this report has been effective in obtaining this objective.
-26-
-------
TECHNICAL REPORT DATA
(Please read instructions on the reverse before completing)
t. REPORT NO.
EPA-600/7-80-080
2.
3. RECIPIENT'S ACCESSION NO.
4, TITLE AND SUBTITLE
A. Case Study in the Use of Ambient Data for
Source Assessment
5. REPORT DATE
April 1980
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Edward T. Brookman and John E. Yocum
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
TRC—The Research Corporation of New England
125 Silas Deane Highway
Wethersfield, Connecticut 06109
10. PROGRAM ELEMENT NO.
INE623
11. CONTRACT/GRANT NO.
68-02-2615, Task 18
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND, PERIOD COVERED
Task Final; 5-8/79
14. SPONSORING AGENCY CODE
EPA/600/13
is. SUPPLEMENTARY NOTES ffiRL-RTP project officer is John O. Milliken, Mail Drop 63,
919/541-2745.
ta. ABSTRACT The report is a case history for an environmental management study to de-
termine what pollution sources are the principal determinants of environmental qual-
ity in Allegheny County, PA. The study was conducted for the Allegheny County
Health Department, Bureau of Air Pollution Control, in Pittsburgh, PA. Results
were used as part of Allegheny County's contribution to the State Implementation
Plan for achieving air quality standards for total suspended particulate matter (TSP).
Techniques used in this ambient-correlation study include: (1) analysis of present
air quality and trends; (2) log-normal distributions; (3) relative frequency of TSP
levels; (4) monthly variations in TSP levels; (5) weekday/weekend analysis; (6) wet-
day/dry-day analysis; (7) analysis of pollution roses; (8) wind frequency analysis;
(9) isopleth maps; (10) contribution of steel plant emissions by modeling; and (11) par-
ticulate identification analysis. The report describes the integrated application of
these techniques to determine the background traditional and nontraditional compo-
nents of the ambient TSP levels. Study results include estimates of the relative
source strengths of the particulates, the relative impacts of the sources, and the
level of confidence of these results.
T7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Pollution
Assessments
Dust
Aerosols
Pollution Control
Stationary Sources
Source Assessment
Particulate
Ambient Data
13B
14B-
11G
07D
13. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
33
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
-27-
------- |